Performance Evaluation of Asynchronous Circuits with Choice Using Abstract Probabilistic Timed Petri Nets

Abstract

A framework for evaluating the performance of asynchronous systems is presented. Due to the dependencies among highly concurrent events performance evaluation of asynchronous circuits is a challenging process. The presented performance model is a Probabilistic Timed Petri-Net (PTPN) with possible choice places to capture the conditional behavior of the system. The proposed framework takes advantage of both static and dynamic analysis to provide precisely enough results in an acceptable time. No data manipulation is done during the simulation phase of the performance evaluation method which leads to very fast simulation. Our proposed performance estimation scheme is faster than usual post-synthesis simulation by an order of 10, while the estimated performance resides in a boundary of 3% of the total imprecision.

High‐Speed Performance Evaluation of a Large‐Scale Shared‐Bus Digital System

“…For this reason, in order to improve performance, the relationship between layout and speed performance must be elucidated . For performance evaluation of digital systems, a method for rapidly calculating the upper bound of speed performance by modeling the system with a stochastic timed Petri net (STPN) and performing iterated Petri net simulation based on the Monte Carlo method has been proposed . Tashiro and colleagues applied this method to large‐scale shared‐bus systems for performance evaluation, but the evaluation time increased sharply as the system scale increased .…”